Recently, Pat Gelsinger, the former CEO of Intel, openly admitted during an industry forum that during Intel’s heyday in the CPU market, its leadership had underestimated the potential of NVIDIA’s GPU technology, viewing it as nothing more than a niche tool for gamers’ graphic rendering. However, with the advent of AI technology, NVIDIA has capitalized on the parallel computing prowess of its GPUs and the robust CUDA ecosystem, securing over 90% of the market share in AI training chips and propelling its market value to a staggering $4.24 trillion. In stark contrast, Intel has witnessed a sharp decline in both market share and stock price, attributed to strategic blunders and technical setbacks, with its current market value now less than a thirtieth of NVIDIA’s. In a bid to reverse its fortunes, Intel has made efforts to catch up by introducing AI chips such as Gaudi3, albeit with limited success. Today, Intel finds itself locked in fierce competition with NVIDIA in the AI computing power arena, while simultaneously collaborating on data center and personal computing products to co-develop new chips, heralding a new shift in the future market landscape.